[tarantool-patches] Re: [PATCH v1 3/4] box: introduce bitmap_majority_test routine

Vladimir Davydov vdavydov.dev at gmail.com
Sat Dec 29 19:16:59 MSK 2018


On Sat, Dec 29, 2018 at 04:57:00PM +0300, Kirill Shcherbatov wrote:
> +/**
> + * Return sizeof(unsigned long)-words aligned size of bitmap by
> + * bit_count - count of bits to set. Size must be aligned, as
> + * bit_sit/bit_clear operations use unsigned long words to setup
> + * bit.
> + */

Changed the comment to

/**
 * @brief Returns the size of memory needed to store a bitmap
 * of \a bit_count bits.
 * The function rounds the size up to a multiple of the word
 * size, which is required by bit_set() and bit_clear().
 * @param bit_count number of bits in the bitmap
 * @retval bitmap size, in bytes
 */

(used doxygen, because all other comments in the file are formated in
the same way)

> +1..5
> +ok 1 - empty bitmap: have 0 expected 0
> +ok 2 - 1-item bitmap: have 8 expected 8
> +ok 3 - 4-items bitmap: have 8 expected 8
> +ok 4 - 64-items bitmap: have 8 expected 8
> +ok 5 - 65-items bitmap: have 16 expected 16
> +	*** test_bitmap: done ***

OK, now I see why you used fail_unless in the previous version - because
all other tests in the file use it. Then you should've told me that
instead of blindly following my instructions - I can make mistakes too.
Changed the test back to fail_unless for consistency and pushed the
patch to 2.1.



More information about the Tarantool-patches mailing list